The urban river network in Tianjin City has stagnant or slow-flow water, which leads to water quality variations different from natural rivers. Multivariate statistical analysis methods, including the cluster analysis method, factor analysis method, and multiple linear regression method, were used to evaluate the water quality data from 11 sites in four main rivers of Tianjin, the Haihe River, Jinhe River, Weijin River, and Waihuai River, for 39 weeks during the period from 2008 to 2009. The results show that the water quality variation exhibited specific temporal and spatial characteristics. Water quality data from May to August were worse than during other months, especially for the ammonia nitrogen and total phosphorous concentrations, which severely exceeded the water quality standards. Water quality deteriorated from the Haihe River to Jinhe River, Weijin River, and, finally, Waihuan River gradually. Nitrogen and phosphorous pollutants were the main pollution sources of the Haihe River, Jinhe River, and Weijin River, while organic pollutants were the main pollution sources of the Waihuan River.